When inducing a tree, one should use the testing performance instead of the validation performance to decide which measure should be used to determine split criteria.

Which of the following statements are false about decision trees? Select one or more. Incorrect answers will be penalized.

If 3 quarters, 1 quarter and none of the instances in a dataset (D) are of target feature classes A, B and C, respectively, then the gini index of D is

Which of the following statements are true about decision tree induction? Select one or more. Incorrect answers will be penalized.

If half, half and none of the instances in a dataset (D) are of target feature classes A, B and C, respectively, then the entropy of D is

The target values of an unbalanced dataset of 10 instances (I1 up to I10) are 0, 1, 1, 1, 0, 1, 0, 1, 1, 1, respectively. An evaluation of a decision tree on this dataset determines that the tree's predictions of the instances I1, I5, I6 and I7 are true negative, false negative, false positive and false negative, respectively. If the accuracy of the decision tree on the dataset is 0.7, what is its precision?